In response to NSF/NIH CRCNS (NS-08-008), we propose a new collaborative project to develop a computational model of the interaction of hippocampus, amygdala, and ventromedial prefrontal cortex in conditioning, extinction, and contextual processing. The model will be applied to data collected from patients with post-traumatic stress disorder (PTSD), which is associated with volumetric and metabolic alterations in these areas. The computational model will also allow us to investigate the possibility that there may be different subtypes of PTSD that involve different nodes of brain dysfunction contributing to a common symptomatology. In addition to PTSD, the computational work may also have future applications to modeling other anxiety disorders and substance abuse disorders (including alcohol abuse and alcoholism) that share many of the same brain substrates as, and that are often co-morbid with, PTSD. ' In parallel with the computational modeling, empirical studies of contextual processing will be conducted in healthy adults and in patients with PTSD, to generate further data to constrain the model, while the model itself will generate new predictions that may drive further empirical studies. The proposed work will increase our understanding of PTSD and examine the idea that it may not be a unified disorder, but a family of pathologies that share features with each other and also with the broader spectrum of anxiety disorders. There will be implications for prevention, through better understanding of pre-existing risk factors, and for optimizing treatment that targets possible PTSD subtypes. Although the empirical work will focus on PTSD, this disorder and alcohol abuse show very high comorbidity, suggesting shared vulnerability. PTSD patients may also use alcohol to self-medicate hyperarousal symptoms. Understanding the interaction between PTSD and alcohol abuse may lead to better treatments for comorbid patients as well as development of prevention strategies targeted to individuals with pre-existing vulnerabilities for developing anxiety disorders vs. chemical dependency vs. both. The project represents a new collaboration among experts on computational neuroscience ofthe hippocampus in conditioning and contextual processing (Myers), on the structural, functional, and behavioral abnormalities in PTSD (Gilbertson, Orr), and on classical fear conditioning in humans and animal models of anxiety (Servatius). RELEVANCE (See instructions): The proposed work will increase our understanding of PTSD, with implications for prevention, through better understanding of pre-existing risk factors, and for optimizing treatment to possible PTSD subtypes. In addition, the computational model of hippocampal-amygdala interaction may have applications to other disorders, particularly alcohol abuse, that share many ofthe same brain substrates, and that may have common pre-existing vulnerabilities; the computational model will provide a framework that can be used to explore these ideas, and may suggest ways to optimize treatment for comorbid patients.